Recently ran through the upgrade process of Fedora 10-11. Preupgrade went smoothly but no upgrade, so had a look at grub.conf and discovered I had to select the upgrade option there during boot. Duly done and complete, kind of.
A $yum update gives me a bunch of dependency issues, here are a couple....
Error: Missing Dependency: libnetfilter_conntrack.so.1 is needed by package iptstate-2.2.1-5.fc11.i586 (installed)
Error: Missing Dependency: libxcb-keysyms.so.0 is needed by package vlc-core-1.0.0-0.11rc3.fc11.i586 (installed)
Error: Missing Dependency: kernel-uname-r = 2.6.27.38-170.2.113.fc10.i686 is needed by package kmod-fglrx-2.6.27.38-170.2.113.fc10.i686-9.3-1.fc10.15.i686 (rpmfusion-nonfree-updates)
You could try using --skip-broken to work around the problem
You could try running: package-cleanup --problems
package-cleanup --dupes
rpm -Va --nofiles --nodigest
Straight away I can see a package is from fc10, so not a big suprise that causes issues. Turns out kmod-fglrx-2.6.27.38-170.2.113.fc10.i686-9.3-1.fc10.15.i686 is from the ATI proprietry drivers I installed, so....
[root@yeti yum.repos.d]# yum remove kmod-fglrx
Loaded plugins: dellsysidplugin2, refresh-packagekit
Setting up Remove Process
Resolving Dependencies
--> Running transaction check
---> Package kmod-fglrx.i686 0:9.3-1.fc10.14 set to be erased
--> Finished Dependency Resolution
...now at least I only have fc11 packages whining at me....
--> Finished Dependency Resolution
iptstate-2.2.1-5.fc11.i586 from installed has depsolving problems
--> Missing Dependency: libnetfilter_conntrack.so.1 is needed by package iptstate-2.2.1-5.fc11.i586 (installed)
vlc-core-1.0.0-0.11rc3.fc11.i586 from installed has depsolving problems
--> Missing Dependency: libxcb-keysyms.so.0 is needed by package vlc-core-1.0.0-0.11rc3.fc11.i586 (installed)
gstreamer-plugins-bad-0.10.11-4.fc11.i586 from installed has depsolving problems
--> Missing Dependency: libass.so.3 is needed by package gstreamer-plugins-bad-0.10.11-4.fc11.i586 (installed)
Error: Missing Dependency: libass.so.3 is needed by package gstreamer-plugins-bad-0.10.11-4.fc11.i586 (installed)
Error: Missing Dependency: libnetfilter_conntrack.so.1 is needed by package iptstate-2.2.1-5.fc11.i586 (installed)
Error: Missing Dependency: libxcb-keysyms.so.0 is needed by package vlc-core-1.0.0-0.11rc3.fc11.i586 (installed)
You could try using --skip-broken to work around the problem
You could try running: package-cleanup --problems
package-cleanup --dupes
rpm -Va --nofiles --nodigest
The gstreamer and VLC packages are packages I want to keep for my video/mp3 collections so I need to work out why these are failing. Clearly (?) they have missing dependacies, but I need to know where to get those from and equally why can I not get them from the repos I have enabled? Dropped from Fedora 11 possibly? In which case I need to find out what replaces them in 11.
Wednesday, 11 November 2009
Monday, 20 July 2009
New Evolution build, still no mobile sync
My hunch anout Evolution recurring appointment bug being the problem seems incorrect. Opened Evolution today, it has been updated to 2.6.3 which fixes that bug.
Tested and indeed I can now create recurring appointments without a crash, but it hasn't sorted my mobile phone sync issues. Back to the drawing board :(
Tested and indeed I can now create recurring appointments without a crash, but it hasn't sorted my mobile phone sync issues. Back to the drawing board :(
Thursday, 16 July 2009
Battling with syncing
Time has come to try to figure out synchronising my mobile with my Fedora desktop calendar program.
For no reason other than it's the default in Fedora, I'm using Evolution as my local calendar program. I have a Nokia N85 and a Google calendar.
Somehow I want to get them all talking/synchronised with each other.
Using open sync and some other stuff (http://www.dafoot.co.uk/index.php?option=com_content&view=article&id=63:fedora-nokia-n85-sync&catid=39:linux&Itemid=66) I successfully got my N85 talking nicely to my FC11 64bit machine, even managed to do a sync of calendar. Once.
More recently I have been consistently getting an error on running the sync tool (see page linked above) rather than a successful sync. Started playing with google calendar, that has a similar but not identical error which makes me think something might have been confused (corrupted) with Evolution.
More so as when I tried to modify a new appointment in Evolution to be weekly with no end date Evolution crashed. And continued to every time I tried to modify that event to be repeating.
Which leads me to think that maybe the sync error is actually evolution getting confused by appoinments in my phone that repeat indefinately. Though that wouldn't fit with not being able to sync with Google unless the Evolution state is so messed up it can't transmit it's data because it got busted on that original (apparently successful) sync.
Seems this bug in Evolution with repeating appts is known and should be fixed in next minor release.... 2.26.3 (I'm on 2.26.2 now). So it's a matter of wait and see now to see if that new version sorts it once it hits the Fedora repositories.
For no reason other than it's the default in Fedora, I'm using Evolution as my local calendar program. I have a Nokia N85 and a Google calendar.
Somehow I want to get them all talking/synchronised with each other.
Using open sync and some other stuff (http://www.dafoot.co.uk/index.php?option=com_content&view=article&id=63:fedora-nokia-n85-sync&catid=39:linux&Itemid=66) I successfully got my N85 talking nicely to my FC11 64bit machine, even managed to do a sync of calendar. Once.
More recently I have been consistently getting an error on running the sync tool (see page linked above) rather than a successful sync. Started playing with google calendar, that has a similar but not identical error which makes me think something might have been confused (corrupted) with Evolution.
More so as when I tried to modify a new appointment in Evolution to be weekly with no end date Evolution crashed. And continued to every time I tried to modify that event to be repeating.
Which leads me to think that maybe the sync error is actually evolution getting confused by appoinments in my phone that repeat indefinately. Though that wouldn't fit with not being able to sync with Google unless the Evolution state is so messed up it can't transmit it's data because it got busted on that original (apparently successful) sync.
Seems this bug in Evolution with repeating appts is known and should be fixed in next minor release.... 2.26.3 (I'm on 2.26.2 now). So it's a matter of wait and see now to see if that new version sorts it once it hits the Fedora repositories.
Wednesday, 1 July 2009
N85 sync with evolution...success :)
Finally got round to looking into getting my Nokia N85 to synchronise its data with my (Fedora) Linux PCs.
After much scrabbling about and fighting with it, it seems to be working now for both my Calendar and my contacts, syncing between the phone and the Evolution Outlook-a-like program.
I've put some notes on the hows and and whys on my personal website:
http://www.dafoot.co.uk/index.php?option=com_content&view=article&id=63:fedora-nokia-n85-sync&catid=39:linux&Itemid=66
Hopefully those notes will prove useful both to me when I get round to doing same to my home machine, but as it runs a slightly different version of Fedora (32 bit FC10) to the machine I've just got work (FC11 64bit) it may be a little more involved!
After much scrabbling about and fighting with it, it seems to be working now for both my Calendar and my contacts, syncing between the phone and the Evolution Outlook-a-like program.
I've put some notes on the hows and and whys on my personal website:
http://www.dafoot.co.uk/index.php?option=com_content&view=article&id=63:fedora-nokia-n85-sync&catid=39:linux&Itemid=66
Hopefully those notes will prove useful both to me when I get round to doing same to my home machine, but as it runs a slightly different version of Fedora (32 bit FC10) to the machine I've just got work (FC11 64bit) it may be a little more involved!
Monday, 29 June 2009
Learning everyday....
Been a while since posting here, learnt a few things during that time.
Booting into runlevel 3 by default can have some odd side effects when you've spent time building scripts in runlevel 5.
Namely, I discovered my backup script wouldn't work for some reason when I tried to fire it off from the command line. I had taken the command generated by GRsync GUI and put it in a #! script so I could run from where ever and not have to remember the command and the flags I was using.
After a bit of poking I determind the backup disk had died. Or so I thought.
I log in locally, startx up x and no disk available. It used to live under /media/backup but it just wasn't there.
Discovered lshw to find out about hardware in the system, that reported the disk was present. So hopefully it was a config thing rather than hardware failure as first feared.
My GRsync (that I had originally used to launch the backup process) wouldn't run either.
Experiment time... logout of X, su and change the run level back to 5 (init 5). Login as regular user and presto, backup works again.
Seems running startx to get to desktop rather than switching run level means some script or other (presumably in /etc/rc5.d) that mounts the backup disk isn't being fired.
Another useful little program I've discovered during investigations is conky. A system monitoring tool. When first launched it is really dull looking black/white, but with the addition of a .conkyrc file to my home area it can be customised very heavily. A quick bit of googling brings up loads of screenshots with sample .conkyrc files to play with.
Been experiencing some strange behaviour with Banshee recently too... when downloading a podcast from the BBC, one one box (FC11 64bit) any - or _ characters in the filename get stripped.
On my home machine (FC10 32bit) the - and _ are left alone and remain in the filename.
On the face of it that wouldn't be an issue so long as each Banshee can find the files right? Problem is I want to setup a process to keep the two postcast directories in sync, so if one or other machine doesn't get turned on the other would make the download. Redundancy if you like to make sure I get the podcast!
Not worked that one out yet :/
Booting into runlevel 3 by default can have some odd side effects when you've spent time building scripts in runlevel 5.
Namely, I discovered my backup script wouldn't work for some reason when I tried to fire it off from the command line. I had taken the command generated by GRsync GUI and put it in a #! script so I could run from where ever and not have to remember the command and the flags I was using.
After a bit of poking I determind the backup disk had died. Or so I thought.
I log in locally, startx up x and no disk available. It used to live under /media/backup but it just wasn't there.
Discovered lshw to find out about hardware in the system, that reported the disk was present. So hopefully it was a config thing rather than hardware failure as first feared.
My GRsync (that I had originally used to launch the backup process) wouldn't run either.
Experiment time... logout of X, su and change the run level back to 5 (init 5). Login as regular user and presto, backup works again.
Seems running startx to get to desktop rather than switching run level means some script or other (presumably in /etc/rc5.d) that mounts the backup disk isn't being fired.
Another useful little program I've discovered during investigations is conky. A system monitoring tool. When first launched it is really dull looking black/white, but with the addition of a .conkyrc file to my home area it can be customised very heavily. A quick bit of googling brings up loads of screenshots with sample .conkyrc files to play with.
Been experiencing some strange behaviour with Banshee recently too... when downloading a podcast from the BBC, one one box (FC11 64bit) any - or _ characters in the filename get stripped.
On my home machine (FC10 32bit) the - and _ are left alone and remain in the filename.
On the face of it that wouldn't be an issue so long as each Banshee can find the files right? Problem is I want to setup a process to keep the two postcast directories in sync, so if one or other machine doesn't get turned on the other would make the download. Redundancy if you like to make sure I get the podcast!
Not worked that one out yet :/
Wednesday, 3 June 2009
Dynamic DNS (contd.)
Turns out I was looking in the wrong place in my router config. I was expecting it to live under the firewall options and open a port there, but instead it is actually under 'Virtual servers'.
Ahh well, either way it works now and I can get CLI access from anywhere to my home machine via SSH :)
Ahh well, either way it works now and I can get CLI access from anywhere to my home machine via SSH :)
Monday, 1 June 2009
Playing with DynDNS
My router is an older one. But it was a good one when I chose it, still does the job even if it can't support latest WiFi encryption tech (so have to rely on WEP!).
What it dows support however is out of the box dynamic DNS. So having a play with it last night and getting some odd results.
If I give my browser the magic URL it goes out, get redirected to my router, which passes the request to a http instance on my LAN (I assume the first one it finds) and passes results back to the browser. So as a proof of concept, I'm happy.
Now the server it is finding is my dev server, which I have not yet put any effort into nailing down http.conf so it displays the default 'welcome to fedora/apache instance' page. Not good.
Simple thinks I, I only want Dynamic DNS for use with SSH so I'll turn on firewall blocking of all ports but the ssh port.
Added to IPTables config on the macine and (re)started IPTables but still the machine seems to be serving out html from httpd :(
Looked at whether I could block incoming requests by port in the router firewall, but not found the options I need yet. I'm sure they should be there somewhere.
Maybe playing with these things at 1 in the morning is not such a good idea :/
Needless to say I switched off DynDNS in my router and PC is now off :P
What it dows support however is out of the box dynamic DNS. So having a play with it last night and getting some odd results.
If I give my browser the magic URL it goes out, get redirected to my router, which passes the request to a http instance on my LAN (I assume the first one it finds) and passes results back to the browser. So as a proof of concept, I'm happy.
Now the server it is finding is my dev server, which I have not yet put any effort into nailing down http.conf so it displays the default 'welcome to fedora/apache instance' page. Not good.
Simple thinks I, I only want Dynamic DNS for use with SSH so I'll turn on firewall blocking of all ports but the ssh port.
Added to IPTables config on the macine and (re)started IPTables but still the machine seems to be serving out html from httpd :(
Looked at whether I could block incoming requests by port in the router firewall, but not found the options I need yet. I'm sure they should be there somewhere.
Maybe playing with these things at 1 in the morning is not such a good idea :/
Needless to say I switched off DynDNS in my router and PC is now off :P
Subscribe to:
Posts (Atom)